Abstract Basic principles are reviewed which provide new access to the synthesis of acentric molecular crystals along with an application to the design of intermolecular interactions and the crystallization of polar host–guest materials. Because of the process of crystal growth, the primary confinement for the alignment of dipolar molecules is given by a surface–nutrient interface and not by the bulk state as usually assumed for the prediction of crystal structures by computational methods. Subject to defined restrictions, spontaneous polarity formation in slowly growing crystals can be regarded as a simple isomerization reaction, ↓ ⇌↑(↓, ↑, orientations of the dipole moments in crystals). Following results of Monte Carlo simulations, performed for a surface layer (adlayer) on a nonrelaxing substrate layer, we conclude the following: (i) Although binding motifs (synthons) are important to induce 1D or 2D order into chains, ribbons, and planes, lateral interactions between such structural elements enter polarity formation by a much higher weight factor. (ii) Particular lateral interactions favoring ↓···↓ are not necessary to obtain polarity in some molecular crystals. Channel-type inclusion materials represent a solution to (ii). Experimental confirmation is provided by, e.g., a large number of polar inclusion compounds of perhydrotriphenylene (PHTP). It is shown that in general a combination of van der Waals interactions for a 2D confinement (alignment of molecular frames) and one of the most nonbonding recognition motifs (↓ vs ↑ orientation) can optimize polarity formation in host–guest lattices. In essence, we review that in some molecular crystals polarity is a tunable property, and that a supramolecular synthesis can produce a material and a property by parallel reactions.